Listening-oriented response generation by exploiting user responses

Jeesoo Bang, Sangdo Han, Jong-Hyeok Lee

Abstract: Highlights • Listening-oriented neural-network-based dialogue system is presented. • Self-disclosure and positiveness are used as listening features. • Listening features of a dialogue are automatically classified. • A user response is inferred from a user input and listening features. • Listening features and a user response are used as conditions in response generation. Abstract Although listening to a conversation partner is a key factor in the success of dialogue systems or conversational agents, recent neural conversation systems have no interest in generating listening-oriented responses. In this paper, we propose an end-to-end dialogue system that generates listening-oriented responses, which make users disclose themselves and feel positive emotions. Our model uses ‘self-disclosure’ and ‘positiveness’ as listening features and generate responses in an appropriate manner to the features. Furthermore, the model infers a user response that will be brought out at the end of the dialogue and uses the inferred user response for generating a system response. By utilizing both listening features and user responses, our model becomes capable of generating listening-oriented responses. In quantitative and qualitative experiments, our model turned out to be capable of generating listening-oriented responses that induce users to disclose themselves and talk positively. The results also show that the model utilizing user responses generates more listening-oriented responses than those only using listening features.